ice: process events created by lag netdev event handler
authorDave Ertman <david.m.ertman@intel.com>
Tue, 20 Jun 2023 22:18:49 +0000 (15:18 -0700)
committerTony Nguyen <anthony.l.nguyen@intel.com>
Thu, 27 Jul 2023 17:56:32 +0000 (10:56 -0700)
Add in the function framework for the processing of LAG events.  Also add
in helper function to perform common tasks.

Add the basis of the process of linking a lower netdev to an upper netdev.

Reviewed-by: Daniel Machon <daniel.machon@microchip.com>
Signed-off-by: Dave Ertman <david.m.ertman@intel.com>
Tested-by: Sujai Buvaneswaran <sujai.buvaneswaran@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
drivers/net/ethernet/intel/ice/ice_lag.c
drivers/net/ethernet/intel/ice/ice_lag.h
drivers/net/ethernet/intel/ice/ice_switch.c
drivers/net/ethernet/intel/ice/ice_switch.h
drivers/net/ethernet/intel/ice/ice_virtchnl.c

index 5725772..ac4deb5 100644 (file)
 #define ICE_LAG_RES_SHARED     BIT(14)
 #define ICE_LAG_RES_VALID      BIT(15)
 
+#define ICE_RECIPE_LEN                 64
+static const u8 ice_dflt_vsi_rcp[ICE_RECIPE_LEN] = {
+       0x05, 0, 0, 0, 0x20, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+       0x85, 0, 0x01, 0, 0, 0, 0xff, 0xff, 0x08, 0, 0, 0, 0, 0, 0, 0,
+       0, 0, 0, 0, 0, 0, 0x30, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
+
 /**
  * ice_lag_set_primary - set PF LAG state as Primary
  * @lag: LAG info struct
@@ -50,6 +57,147 @@ static void ice_lag_set_backup(struct ice_lag *lag)
        lag->role = ICE_LAG_BACKUP;
 }
 
+/**
+ * ice_netdev_to_lag - return pointer to associated lag struct from netdev
+ * @netdev: pointer to net_device struct to query
+ */
+static struct ice_lag *ice_netdev_to_lag(struct net_device *netdev)
+{
+       struct ice_netdev_priv *np;
+       struct ice_vsi *vsi;
+
+       if (!netif_is_ice(netdev))
+               return NULL;
+
+       np = netdev_priv(netdev);
+       if (!np)
+               return NULL;
+
+       vsi = np->vsi;
+       if (!vsi)
+               return NULL;
+
+       return vsi->back->lag;
+}
+
+/**
+ * ice_lag_find_primary - returns pointer to primary interfaces lag struct
+ * @lag: local interfaces lag struct
+ */
+static struct ice_lag *ice_lag_find_primary(struct ice_lag *lag)
+{
+       struct ice_lag *primary_lag = NULL;
+       struct list_head *tmp;
+
+       list_for_each(tmp, lag->netdev_head) {
+               struct ice_lag_netdev_list *entry;
+               struct ice_lag *tmp_lag;
+
+               entry = list_entry(tmp, struct ice_lag_netdev_list, node);
+               tmp_lag = ice_netdev_to_lag(entry->netdev);
+               if (tmp_lag && tmp_lag->primary) {
+                       primary_lag = tmp_lag;
+                       break;
+               }
+       }
+
+       return primary_lag;
+}
+
+/**
+ * ice_lag_cfg_dflt_fltr - Add/Remove default VSI rule for LAG
+ * @lag: lag struct for local interface
+ * @add: boolean on whether we are adding filters
+ */
+static int
+ice_lag_cfg_dflt_fltr(struct ice_lag *lag, bool add)
+{
+       struct ice_sw_rule_lkup_rx_tx *s_rule;
+       u16 s_rule_sz, vsi_num;
+       struct ice_hw *hw;
+       u32 act, opc;
+       u8 *eth_hdr;
+       int err;
+
+       hw = &lag->pf->hw;
+       vsi_num = ice_get_hw_vsi_num(hw, 0);
+
+       s_rule_sz = ICE_SW_RULE_RX_TX_ETH_HDR_SIZE(s_rule);
+       s_rule = kzalloc(s_rule_sz, GFP_KERNEL);
+       if (!s_rule) {
+               dev_err(ice_pf_to_dev(lag->pf), "error allocating rule for LAG default VSI\n");
+               return -ENOMEM;
+       }
+
+       if (add) {
+               eth_hdr = s_rule->hdr_data;
+               ice_fill_eth_hdr(eth_hdr);
+
+               act = (vsi_num << ICE_SINGLE_ACT_VSI_ID_S) &
+                       ICE_SINGLE_ACT_VSI_ID_M;
+               act |= ICE_SINGLE_ACT_VSI_FORWARDING |
+                       ICE_SINGLE_ACT_VALID_BIT | ICE_SINGLE_ACT_LAN_ENABLE;
+
+               s_rule->hdr.type = cpu_to_le16(ICE_AQC_SW_RULES_T_LKUP_RX);
+               s_rule->recipe_id = cpu_to_le16(lag->pf_recipe);
+               s_rule->src = cpu_to_le16(hw->port_info->lport);
+               s_rule->act = cpu_to_le32(act);
+               s_rule->hdr_len = cpu_to_le16(DUMMY_ETH_HDR_LEN);
+               opc = ice_aqc_opc_add_sw_rules;
+       } else {
+               s_rule->index = cpu_to_le16(lag->pf_rule_id);
+               opc = ice_aqc_opc_remove_sw_rules;
+       }
+
+       err = ice_aq_sw_rules(&lag->pf->hw, s_rule, s_rule_sz, 1, opc, NULL);
+       if (err)
+               goto dflt_fltr_free;
+
+       if (add)
+               lag->pf_rule_id = le16_to_cpu(s_rule->index);
+       else
+               lag->pf_rule_id = 0;
+
+dflt_fltr_free:
+       kfree(s_rule);
+       return err;
+}
+
+/**
+ * ice_lag_cfg_pf_fltrs - set filters up for new active port
+ * @lag: local interfaces lag struct
+ * @ptr: opaque data containing notifier event
+ */
+static void
+ice_lag_cfg_pf_fltrs(struct ice_lag *lag, void *ptr)
+{
+       struct netdev_notifier_bonding_info *info;
+       struct netdev_bonding_info *bonding_info;
+       struct net_device *event_netdev;
+       struct device *dev;
+
+       event_netdev = netdev_notifier_info_to_dev(ptr);
+       /* not for this netdev */
+       if (event_netdev != lag->netdev)
+               return;
+
+       info = (struct netdev_notifier_bonding_info *)ptr;
+       bonding_info = &info->bonding_info;
+       dev = ice_pf_to_dev(lag->pf);
+
+       /* interface not active - remove old default VSI rule */
+       if (bonding_info->slave.state && lag->pf_rule_id) {
+               if (ice_lag_cfg_dflt_fltr(lag, false))
+                       dev_err(dev, "Error removing old default VSI filter\n");
+               return;
+       }
+
+       /* interface becoming active - add new default VSI rule */
+       if (!bonding_info->slave.state && !lag->pf_rule_id)
+               if (ice_lag_cfg_dflt_fltr(lag, true))
+                       dev_err(dev, "Error adding new default VSI filter\n");
+}
+
 /**
  * ice_display_lag_info - print LAG info
  * @lag: LAG info struct
@@ -85,6 +233,83 @@ static void ice_display_lag_info(struct ice_lag *lag)
                upper, role, primary);
 }
 
+/**
+ * ice_lag_move_vf_node_tc - move scheduling nodes for one VF on one TC
+ * @lag: lag info struct
+ * @oldport: lport of previous nodes location
+ * @newport: lport of destination nodes location
+ * @vsi_num: array index of VSI in PF space
+ * @tc: traffic class to move
+ */
+static void
+ice_lag_move_vf_node_tc(struct ice_lag *lag, u8 oldport, u8 newport,
+                       u16 vsi_num, u8 tc)
+{
+}
+
+/**
+ * ice_lag_move_single_vf_nodes - Move Tx scheduling nodes for single VF
+ * @lag: primary interface LAG struct
+ * @oldport: lport of previous interface
+ * @newport: lport of destination interface
+ * @vsi_num: SW index of VF's VSI
+ */
+static void
+ice_lag_move_single_vf_nodes(struct ice_lag *lag, u8 oldport, u8 newport,
+                            u16 vsi_num)
+{
+       u8 tc;
+
+       ice_for_each_traffic_class(tc)
+               ice_lag_move_vf_node_tc(lag, oldport, newport, vsi_num, tc);
+}
+
+/**
+ * ice_lag_move_new_vf_nodes - Move Tx scheduling nodes for a VF if required
+ * @vf: the VF to move Tx nodes for
+ *
+ * Called just after configuring new VF queues. Check whether the VF Tx
+ * scheduling nodes need to be updated to fail over to the active port. If so,
+ * move them now.
+ */
+void ice_lag_move_new_vf_nodes(struct ice_vf *vf)
+{
+}
+
+/**
+ * ice_lag_move_vf_nodes - move Tx scheduling nodes for all VFs to new port
+ * @lag: lag info struct
+ * @oldport: lport of previous interface
+ * @newport: lport of destination interface
+ */
+static void ice_lag_move_vf_nodes(struct ice_lag *lag, u8 oldport, u8 newport)
+{
+       struct ice_pf *pf;
+       int i;
+
+       if (!lag->primary)
+               return;
+
+       pf = lag->pf;
+       ice_for_each_vsi(pf, i)
+               if (pf->vsi[i] && (pf->vsi[i]->type == ICE_VSI_VF ||
+                                  pf->vsi[i]->type == ICE_VSI_SWITCHDEV_CTRL))
+                       ice_lag_move_single_vf_nodes(lag, oldport, newport, i);
+}
+
+#define ICE_LAG_SRIOV_CP_RECIPE                10
+#define ICE_LAG_SRIOV_TRAIN_PKT_LEN    16
+
+/**
+ * ice_lag_cfg_cp_fltr - configure filter for control packets
+ * @lag: local interface's lag struct
+ * @add: add or remove rule
+ */
+static void
+ice_lag_cfg_cp_fltr(struct ice_lag *lag, bool add)
+{
+}
+
 /**
  * ice_lag_info_event - handle NETDEV_BONDING_INFO event
  * @lag: LAG info struct
@@ -126,105 +351,168 @@ lag_out:
        ice_display_lag_info(lag);
 }
 
+/**
+ * ice_lag_reclaim_vf_tc - move scheduling nodes back to primary interface
+ * @lag: primary interface lag struct
+ * @src_hw: HW struct current node location
+ * @vsi_num: VSI index in PF space
+ * @tc: traffic class to move
+ */
+static void
+ice_lag_reclaim_vf_tc(struct ice_lag *lag, struct ice_hw *src_hw, u16 vsi_num,
+                     u8 tc)
+{
+}
+
+/**
+ * ice_lag_reclaim_vf_nodes - When interface leaving bond primary reclaims nodes
+ * @lag: primary interface lag struct
+ * @src_hw: HW struct for current node location
+ */
+static void
+ice_lag_reclaim_vf_nodes(struct ice_lag *lag, struct ice_hw *src_hw)
+{
+       struct ice_pf *pf;
+       int i, tc;
+
+       if (!lag->primary || !src_hw)
+               return;
+
+       pf = lag->pf;
+       ice_for_each_vsi(pf, i)
+               if (pf->vsi[i] && (pf->vsi[i]->type == ICE_VSI_VF ||
+                                  pf->vsi[i]->type == ICE_VSI_SWITCHDEV_CTRL))
+                       ice_for_each_traffic_class(tc)
+                               ice_lag_reclaim_vf_tc(lag, src_hw, i, tc);
+}
+
 /**
  * ice_lag_link - handle LAG link event
  * @lag: LAG info struct
- * @info: info from the netdev notifier
  */
-static void
-ice_lag_link(struct ice_lag *lag, struct netdev_notifier_changeupper_info *info)
+static void ice_lag_link(struct ice_lag *lag)
 {
-       struct net_device *netdev_tmp, *upper = info->upper_dev;
        struct ice_pf *pf = lag->pf;
-       int peers = 0;
 
        if (lag->bonded)
                dev_warn(ice_pf_to_dev(pf), "%s Already part of a bond\n",
                         netdev_name(lag->netdev));
 
-       rcu_read_lock();
-       for_each_netdev_in_bond_rcu(upper, netdev_tmp)
-               peers++;
-       rcu_read_unlock();
-
-       if (lag->upper_netdev != upper) {
-               dev_hold(upper);
-               lag->upper_netdev = upper;
-       }
-
-       ice_clear_rdma_cap(pf);
-
        lag->bonded = true;
        lag->role = ICE_LAG_UNSET;
-
-       /* if this is the first element in an LAG mark as primary */
-       lag->primary = !!(peers == 1);
 }
 
 /**
  * ice_lag_unlink - handle unlink event
  * @lag: LAG info struct
- * @info: info from netdev notification
  */
-static void
-ice_lag_unlink(struct ice_lag *lag,
-              struct netdev_notifier_changeupper_info *info)
+static void ice_lag_unlink(struct ice_lag *lag)
 {
-       struct net_device *netdev_tmp, *upper = info->upper_dev;
+       u8 pri_port, act_port, loc_port;
        struct ice_pf *pf = lag->pf;
-       bool found = false;
 
        if (!lag->bonded) {
                netdev_dbg(lag->netdev, "bonding unlink event on non-LAG netdev\n");
                return;
        }
 
-       /* determine if we are in the new LAG config or not */
-       rcu_read_lock();
-       for_each_netdev_in_bond_rcu(upper, netdev_tmp) {
-               if (netdev_tmp == lag->netdev) {
-                       found = true;
-                       break;
+       if (lag->primary) {
+               act_port = lag->active_port;
+               pri_port = lag->pf->hw.port_info->lport;
+               if (act_port != pri_port && act_port != ICE_LAG_INVALID_PORT)
+                       ice_lag_move_vf_nodes(lag, act_port, pri_port);
+               lag->primary = false;
+               lag->active_port = ICE_LAG_INVALID_PORT;
+       } else {
+               struct ice_lag *primary_lag;
+
+               primary_lag = ice_lag_find_primary(lag);
+               if (primary_lag) {
+                       act_port = primary_lag->active_port;
+                       pri_port = primary_lag->pf->hw.port_info->lport;
+                       loc_port = pf->hw.port_info->lport;
+                       if (act_port == loc_port &&
+                           act_port != ICE_LAG_INVALID_PORT) {
+                               ice_lag_reclaim_vf_nodes(primary_lag,
+                                                        &lag->pf->hw);
+                               primary_lag->active_port = ICE_LAG_INVALID_PORT;
+                       }
                }
        }
-       rcu_read_unlock();
-
-       if (found)
-               return;
-
-       if (lag->upper_netdev) {
-               dev_put(lag->upper_netdev);
-               lag->upper_netdev = NULL;
-       }
 
-       ice_set_rdma_cap(pf);
        lag->bonded = false;
        lag->role = ICE_LAG_NONE;
+       lag->upper_netdev = NULL;
 }
 
 /**
- * ice_lag_unregister - handle netdev unregister events
- * @lag: LAG info struct
- * @netdev: netdev reporting the event
+ * ice_lag_link_unlink - helper function to call lag_link/unlink
+ * @lag: lag info struct
+ * @ptr: opaque pointer data
  */
-static void ice_lag_unregister(struct ice_lag *lag, struct net_device *netdev)
+static void ice_lag_link_unlink(struct ice_lag *lag, void *ptr)
 {
-       struct ice_pf *pf = lag->pf;
+       struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
+       struct netdev_notifier_changeupper_info *info = ptr;
 
-       /* check to see if this event is for this netdev
-        * check that we are in an aggregate
-        */
-       if (netdev != lag->netdev || !lag->bonded)
+       if (netdev != lag->netdev)
                return;
 
-       if (lag->upper_netdev) {
-               dev_put(lag->upper_netdev);
-               lag->upper_netdev = NULL;
-               ice_set_rdma_cap(pf);
-       }
-       /* perform some cleanup in case we come back */
-       lag->bonded = false;
-       lag->role = ICE_LAG_NONE;
+       if (info->linking)
+               ice_lag_link(lag);
+       else
+               ice_lag_unlink(lag);
+}
+
+/**
+ * ice_lag_set_swid - set the SWID on secondary interface
+ * @primary_swid: primary interface's SWID
+ * @local_lag: local interfaces LAG struct
+ * @link: Is this a linking activity
+ *
+ * If link is false, then primary_swid should be expected to not be valid
+ */
+static void
+ice_lag_set_swid(u16 primary_swid, struct ice_lag *local_lag,
+                bool link)
+{
+}
+
+/**
+ * ice_lag_primary_swid - set/clear the SHARED attrib of primary's SWID
+ * @lag: primary interfaces lag struct
+ * @link: is this a linking activity
+ *
+ * Implement setting primary SWID as shared using 0x020B
+ */
+static void ice_lag_primary_swid(struct ice_lag *lag, bool link)
+{
+       struct ice_hw *hw;
+       u16 swid;
+
+       hw = &lag->pf->hw;
+       swid = hw->port_info->sw_id;
+
+       if (ice_share_res(hw, ICE_AQC_RES_TYPE_SWID, link, swid))
+               dev_warn(ice_pf_to_dev(lag->pf), "Failure to set primary interface shared status\n");
+}
+
+/**
+ * ice_lag_add_prune_list - Adds event_pf's VSI to primary's prune list
+ * @lag: lag info struct
+ * @event_pf: PF struct for VSI we are adding to primary's prune list
+ */
+static void ice_lag_add_prune_list(struct ice_lag *lag, struct ice_pf *event_pf)
+{
+}
+
+/**
+ * ice_lag_del_prune_list - Remove secondary's vsi from primary's prune list
+ * @lag: primary interface's ice_lag struct
+ * @event_pf: PF struct for unlinking interface
+ */
+static void ice_lag_del_prune_list(struct ice_lag *lag, struct ice_pf *event_pf)
+{
 }
 
 /**
@@ -257,6 +545,7 @@ static void ice_lag_init_feature_support_flag(struct ice_pf *pf)
 static void ice_lag_changeupper_event(struct ice_lag *lag, void *ptr)
 {
        struct netdev_notifier_changeupper_info *info;
+       struct ice_lag *primary_lag;
        struct net_device *netdev;
 
        info = ptr;
@@ -266,24 +555,107 @@ static void ice_lag_changeupper_event(struct ice_lag *lag, void *ptr)
        if (netdev != lag->netdev)
                return;
 
-       if (!info->upper_dev) {
-               netdev_dbg(netdev, "changeupper rcvd, but no upper defined\n");
-               return;
+       primary_lag = ice_lag_find_primary(lag);
+       if (info->linking) {
+               lag->upper_netdev = info->upper_dev;
+               /* If there is not already a primary interface in the LAG,
+                * then mark this one as primary.
+                */
+               if (!primary_lag) {
+                       lag->primary = true;
+                       /* Configure primary's SWID to be shared */
+                       ice_lag_primary_swid(lag, true);
+                       primary_lag = lag;
+               } else {
+                       u16 swid;
+
+                       swid = primary_lag->pf->hw.port_info->sw_id;
+                       ice_lag_set_swid(swid, lag, true);
+                       ice_lag_add_prune_list(primary_lag, lag->pf);
+               }
+               /* add filter for primary control packets */
+               ice_lag_cfg_cp_fltr(lag, true);
+       } else {
+               if (!primary_lag && lag->primary)
+                       primary_lag = lag;
+
+               if (primary_lag) {
+                       if (!lag->primary) {
+                               ice_lag_set_swid(0, lag, false);
+                       } else {
+                               ice_lag_primary_swid(lag, false);
+                               ice_lag_del_prune_list(primary_lag, lag->pf);
+                       }
+                       ice_lag_cfg_cp_fltr(lag, false);
+               }
        }
+}
+
+/**
+ * ice_lag_monitor_link - monitor interfaces entering/leaving the aggregate
+ * @lag: lag info struct
+ * @ptr: opaque data containing notifier event
+ *
+ * This function only operates after a primary has been set.
+ */
+static void ice_lag_monitor_link(struct ice_lag *lag, void *ptr)
+{
+}
+
+/**
+ * ice_lag_monitor_active - main PF keep track of which port is active
+ * @lag: lag info struct
+ * @ptr: opaque data containing notifier event
+ *
+ * This function is for the primary PF to monitor changes in which port is
+ * active and handle changes for SRIOV VF functionality
+ */
+static void ice_lag_monitor_active(struct ice_lag *lag, void *ptr)
+{
+}
+
+/**
+ * ice_lag_chk_comp - evaluate bonded interface for feature support
+ * @lag: lag info struct
+ * @ptr: opaque data for netdev event info
+ */
+static bool
+ice_lag_chk_comp(struct ice_lag *lag, void *ptr)
+{
+       return true;
+}
+
+/**
+ * ice_lag_unregister - handle netdev unregister events
+ * @lag: LAG info struct
+ * @event_netdev: netdev struct for target of notifier event
+ */
+static void
+ice_lag_unregister(struct ice_lag *lag, struct net_device *event_netdev)
+{
+}
 
-       netdev_dbg(netdev, "bonding %s\n", info->linking ? "LINK" : "UNLINK");
+/**
+ * ice_lag_monitor_rdma - set and clear rdma functionality
+ * @lag: pointer to lag struct
+ * @ptr: opaque data for netdev event info
+ */
+static void
+ice_lag_monitor_rdma(struct ice_lag *lag, void *ptr)
+{
+       struct netdev_notifier_changeupper_info *info;
+       struct net_device *netdev;
+
+       info = ptr;
+       netdev = netdev_notifier_info_to_dev(ptr);
 
-       if (!netif_is_lag_master(info->upper_dev)) {
-               netdev_dbg(netdev, "changeupper rcvd, but not primary. bail\n");
+       if (netdev != lag->netdev)
                return;
-       }
 
        if (info->linking)
-               ice_lag_link(lag, info);
+               ice_clear_rdma_cap(lag->pf);
        else
-               ice_lag_unlink(lag, info);
-
-       ice_display_lag_info(lag);
+               ice_set_rdma_cap(lag->pf);
 }
 
 /**
@@ -307,16 +679,31 @@ static void ice_lag_process_event(struct work_struct *work)
        switch (lag_work->event) {
        case NETDEV_CHANGEUPPER:
                info = &lag_work->info.changeupper_info;
-               ice_lag_changeupper_event(lag_work->lag, info);
+               if (ice_is_feature_supported(pf, ICE_F_SRIOV_LAG)) {
+                       ice_lag_monitor_link(lag_work->lag, info);
+                       ice_lag_changeupper_event(lag_work->lag, info);
+                       ice_lag_link_unlink(lag_work->lag, info);
+               }
+               ice_lag_monitor_rdma(lag_work->lag, info);
                break;
        case NETDEV_BONDING_INFO:
+               if (ice_is_feature_supported(pf, ICE_F_SRIOV_LAG)) {
+                       if (!ice_lag_chk_comp(lag_work->lag,
+                                             &lag_work->info.bonding_info)) {
+                               goto lag_cleanup;
+                       }
+                       ice_lag_monitor_active(lag_work->lag,
+                                              &lag_work->info.bonding_info);
+                       ice_lag_cfg_pf_fltrs(lag_work->lag,
+                                            &lag_work->info.bonding_info);
+               }
                ice_lag_info_event(lag_work->lag, &lag_work->info.bonding_info);
                break;
        case NETDEV_UNREGISTER:
                if (ice_is_feature_supported(pf, ICE_F_SRIOV_LAG)) {
                        netdev = lag_work->info.bonding_info.info.dev;
-                       if ((netdev == lag_work->lag->netdev ||
-                            lag_work->lag->primary) && lag_work->lag->bonded)
+                       if (netdev == lag_work->lag->netdev &&
+                           lag_work->lag->bonded)
                                ice_lag_unregister(lag_work->lag, netdev);
                }
                break;
@@ -324,6 +711,7 @@ static void ice_lag_process_event(struct work_struct *work)
                break;
        }
 
+lag_cleanup:
        /* cleanup resources allocated for this work item */
        list_for_each_safe(tmp, n, &lag_work->netdev_list.node) {
                struct ice_lag_netdev_list *entry;
@@ -465,6 +853,22 @@ static void ice_unregister_lag_handler(struct ice_lag *lag)
        }
 }
 
+/**
+ * ice_create_lag_recipe
+ * @hw: pointer to HW struct
+ * @base_recipe: recipe to base the new recipe on
+ * @prio: priority for new recipe
+ *
+ * function returns 0 on error
+ */
+static u16 ice_create_lag_recipe(struct ice_hw *hw, const u8 *base_recipe,
+                                u8 prio)
+{
+       u16 rid = 0;
+
+       return rid;
+}
+
 /**
  * ice_init_lag - initialize support for LAG
  * @pf: PF struct
@@ -506,6 +910,12 @@ int ice_init_lag(struct ice_pf *pf)
                goto lag_error;
        }
 
+       lag->pf_recipe = ice_create_lag_recipe(&pf->hw, ice_dflt_vsi_rcp, 1);
+       if (!lag->pf_recipe) {
+               err = -EINVAL;
+               goto lag_error;
+       }
+
        ice_display_lag_info(lag);
 
        dev_dbg(dev, "INIT LAG complete\n");
@@ -538,6 +948,9 @@ void ice_deinit_lag(struct ice_pf *pf)
 
        flush_workqueue(ice_lag_wq);
 
+       ice_free_hw_res(&pf->hw, ICE_AQC_RES_TYPE_RECIPE, 1,
+                       &pf->lag->pf_recipe);
+
        kfree(lag);
 
        pf->lag = NULL;
index a4314a4..df4af51 100644 (file)
@@ -56,6 +56,7 @@ struct ice_lag_work {
        } info;
 };
 
+void ice_lag_move_new_vf_nodes(struct ice_vf *vf);
 int ice_init_lag(struct ice_pf *pf);
 void ice_deinit_lag(struct ice_pf *pf);
 #endif /* _ICE_LAG_H_ */
index 8c332d2..91bc92f 100644 (file)
@@ -25,7 +25,6 @@
  *      In case of Ether type filter it is treated as header without VLAN tag
  *      and byte 12 and 13 is used to program a given Ether type instead
  */
-#define DUMMY_ETH_HDR_LEN              16
 static const u8 dummy_eth_header[DUMMY_ETH_HDR_LEN] = { 0x2, 0, 0, 0, 0, 0,
                                                        0x2, 0, 0, 0, 0, 0,
                                                        0x81, 0, 0, 0};
@@ -2460,6 +2459,15 @@ static void ice_fill_sw_info(struct ice_hw *hw, struct ice_fltr_info *fi)
        }
 }
 
+/**
+ * ice_fill_eth_hdr - helper to copy dummy_eth_hdr into supplied buffer
+ * @eth_hdr: pointer to buffer to populate
+ */
+void ice_fill_eth_hdr(u8 *eth_hdr)
+{
+       memcpy(eth_hdr, dummy_eth_header, DUMMY_ETH_HDR_LEN);
+}
+
 /**
  * ice_fill_sw_rule - Helper function to fill switch rule structure
  * @hw: pointer to the hardware structure
index 51cd149..250823a 100644 (file)
@@ -30,6 +30,8 @@
        ICE_SW_RULE_RX_TX_HDR_SIZE((s), 0)
 #define ICE_SW_RULE_LG_ACT_SIZE(s, n)          struct_size((s), act, (n))
 
+#define DUMMY_ETH_HDR_LEN              16
+
 /* VSI context structure for add/get/update/free operations */
 struct ice_vsi_ctx {
        u16 vsi_num;
@@ -402,6 +404,7 @@ u16 ice_get_hw_vsi_num(struct ice_hw *hw, u16 vsi_handle);
 
 int ice_replay_vsi_all_fltr(struct ice_hw *hw, u16 vsi_handle);
 void ice_rm_all_sw_replay_rule_info(struct ice_hw *hw);
+void ice_fill_eth_hdr(u8 *eth_hdr);
 
 int
 ice_aq_sw_rules(struct ice_hw *hw, void *rule_list, u16 rule_list_sz,
index efbc296..625da88 100644 (file)
@@ -1724,6 +1724,8 @@ error_param:
                                vf->vf_id, i);
        }
 
+       ice_lag_move_new_vf_nodes(vf);
+
        /* send the response to the VF */
        return ice_vc_send_msg_to_vf(vf, VIRTCHNL_OP_CONFIG_VSI_QUEUES,
                                     VIRTCHNL_STATUS_ERR_PARAM, NULL, 0);